The Roman city of Calleva Atrebatum (Silchester, Hampshire) was subjected to an extensive programme of excavations by the Society of Antiquaries of London between 1890 and 1909. The turn of insula IX came in 1893-4, when one of the most important discoveries was a large town house (House 1) whose north-east/south-west orientation was completely different from that of the Roman street grid aligned on the cardinal points (Fox 1895; Fulford and Clarke 2002; Clarke et al. 2001). Apart from the observation that a well containing a dwarf Roman column incised with a name in ogham script could only ‘have been sunk when the portion of the house in which it occurred was in a ruinous condition’ (Fox 1895, 441), no other comment was made about the dating or development of the house.

Location of Silchester and insula IX in the context of the Society of Antiquaries' plan of buildings within the walls at the conclusion of their excavation of Silchester in 1908
Location of Silchester and insula IX in the context of the Society of Antiquaries' plan of buildings within the walls at the conclusion of their excavation of Silchester in 1908.

In 1997 a new programme of excavations began on part of insula IX with the aim of exploring the full sequence of occupation from the late Iron Age through to post-Roman abandonment (Clarke and Fulford 2002). It rapidly emerged that ‘House 1’ had a complex history of development, with abandonment and demolition in the mid 3rd century AD. The aim of this article is to report that development from its origins around the mid-1st century AD to demolition about the middle of the 3rd century AD.

Victorian excavations Plan
Left: Plan of the Victorian excavations of insula IX, 1893-4 (after Archaeologia 54 (2), pl. XLV)
Right: Plan of insula IX showing the area under excavation since 1997, the remains found in 1893-4 and the plans of buildings plotted from aerial photography by RCHM


The approach to the publication of excavation reports, particularly since the great increase in information arising from the closer observation and recording of archaeological stratigraphy and from the reporting of associated material culture and biological data has been to provide a synthesis and interpretation of the excavated sequence and the associated finds evidence. Initially, it was possible to provide supplementary information on stratigraphy and finds through microfiche. Latterly electronically generated databases with detailed information about excavation and finds have been made available on-line, very largely through the services of ADS/AHDS, to complement the printed report. Nevertheless for most large projects, even in the era of microfiche, the possibility of testing interpretations and conclusions has rested on the re-examination of the original site and finds archive wherever they were archived.

From the outset of the insula IX project the plan, context record and basic finds data were entered onto a database (IADB). The excavation photography soon followed. As research on the finds developed, for example on coins and other ‘small finds’, that expert information was also entered. The realisation that the insula IX project was going to take many seasons of fieldwork and years of post-excavation analysis prompted a policy to, first, publish in appropriate form while the fieldwork continued and, second, to link text-rich, printed articles and reports of the excavation with associated websites where the primary evidence could be scrutinised on-line. Our first pilot was to report on the Victorian excavation of insula IX in 1893-4 (Clarke et al. 2001; Fulford and Clarke 2002). While the printed article contained a summary of the field and finds evidence with limited photo reproduction, the website allowed the reader to search through to the records of each relevant context and plan record, supported by the full archive of digitised colour photography. Information on the Victorian finds was complemented with a much richer array of images than was possible in the journal article. There was over a year's interval between the electronic and the printed publication.

Much more substantial was the second phase of reporting on the excavation. This involved a large body of stratigraphic and finds information concerning the late Roman occupation of the insula in the period from AD 250/300 to the abandonment in the 6th-7th century. In addition to the context and plan record there was a wide range of artefacts and biological data, with some categories containing thousands of items of information. The Late Roman website was launched in the spring of 2005 (Clarke et al. 2005), while the complementary printed volume was published in late 2006 (Fulford et al. 2006). The website gives access to all the raw data associated with the stratigraphic and finds records, while the printed volume concentrates on the text-rich synthesis, analysis and discussion of results. However, it was decided at the outset that both publications would need to stand in their own right. Without some interpretive structure the user of the website would find it hard to navigate through the underlying archive; equally, the user of the printed report would require a fairly detailed narrative of the excavation and tabulated data to support specialist reports in order to understand the significance of the stratigraphic and structural sequence and their associated finds. As a result there is some overlap in content between the two. In particular the interpretation and synthesis of the stratigraphy on the website, which leads the reader into the related archive, is reproduced in the printed report. In this article, however, all duplication is avoided. The description of the sequence of buildings which occupy the ‘House 1’ plot tries to avoid detailed description which can be pursued through the linked archive and to focus on interpretation.

Introduction to the ‘House 1’ site and sequence

The masonry remains of the House 1 described by Fox (1895) proved, on excavation, to comprise the remains of two successive periods of building, the earlier of which consisted of the remains of two detached houses (our Period 3 Masonry Buildings 1 and 2). The later comprised the remains of one large, possibly aisled, town-house (our Period 4 Masonry Building 3). In addition, and beneath the remains of the Early Roman masonry buildings, were the remains of a large, timber-framed house (our Period 2 Timber Building 2) and a circular building (Period 2 Timber Building 3). To the north-east, the ground plan extended beyond the limits of the Period 4 Masonry Building 3. Immediately adjacent to it, and extending its line, was a further, detached timber building (Period 2 Timber Building 1), which occupied the north-east angle of the insula. Although the latter falls outside the footprint of Period 4, Masonry Building 3, it has been included here as it appears to be intimately connected with the life of the Period 2, Timber Building 2).

With the exception of our earliest Period 1, all buildings are later than the imposition of the orthogonal street grid laid out on the cardinal points. The unusual north-east/south-west orientation established for the Period 2 buildings was maintained through to the demolition of the Period 4 House 3 in the mid-3rd century AD. This report focuses only on the sequence of buildings which occupied the same, House 1, site from c. AD 70-80 within the excavation area. The wider, period-by-period, developments of the archaeology within the excavation area will be reported in subsequent publications. The excavation is still continuing and this will shed further light on Period 1 and the earliest phases of the Period 2 Timber Buildings.

Disturbance of the stratigraphy

While the stratigraphy associated with the latest phase, Period 4 Masonry Building 3, has been truncated or disturbed both by the digging of late Roman pits through its walls and floors, and the excavation of 1893, that of the earlier buildings has been additionally disturbed by the foundation digging of their successors. The wall-following excavation of 1893 has completely isolated the stratigraphy of the two periods of masonry building from that of the rest of the insula, while the combination of Victorian excavation and foundation-trenching for the Roman masonry buildings has had a further impact on the underlying timber buildings. Victorian trial-trenching, particularly around its central hearth, has similarly disturbed the stratigraphy associated with the Period 2 Timber Building 1 and its successor.

The archive

Throughout this report there are links which will take the reader to the underlying archive of the excavation and supporting documentation concerning the specialists' reports on the finds. The archive for this report, particularly in terms of context and finds' records, dates back to the earliest seasons of excavation. Interpretations, both at the level of individual contexts and of larger groups of contexts (Objects), such as buildings, have changed over the years. Research of the archive will, therefore, give the reader a sense of how interpretations have changed over time. No attempt has been made to revise all records to be consistent with the interpretations presented here.

The Objects

The contexts and the stratigraphic relations associated with each building and period of building have been grouped together as Objects with their associated stratigraphic matrices; and both can be searched in the database held by the Archaeology Data Service:

Period 1: c. AD 40-50 – c. AD 70-80

Period 2: c. AD 70-80 – c. AD 125-50

Period 3: c. AD 125-50 – c. AD 200

Period 4: c. AD 200 – c. AD 250


© Internet Archaeology URL:
Last updated: Wed Sept 12 2007